How Not to Network a Nation. The Uneasy History of the Soviet Internet

(Ben Green) #1

100 Chapter 3

society processes can be expressed as an exponential function of the transis-
tors on the circuits its industries can produce.^52 These sibling laws (Moore’s
2 N and Kharkevich’s N^2 ) diverge interestingly in complex systems (when N
is larger than 4). They also backlight their micro and macro focuses—Moore
on microscopic industrial production and Kharkevich on informational
industrial society. The result is different framings of the national network
as a sort of central processor. Like Baran, Kharkevich prioritizes building
“survivable” military networks while also looking to benefit other civilian
and social goals.
Unlike the cybernetic metaphor of the brain that Baran drew from War-
ren McCulloch, Kharkevich saw his national network as a nervous system
that was overlaid onto the body of the nation and that would be governed
by a central processor, or brain, located in Moscow. The resulting contrast
of cybernetic metaphors for the information societies is again sharp: for
Kharkevich, the networked nation was the body controlled by a central
brain, and for Baran, the networked nation was the brain itself.
Like other network designers, Kharkevich also designed the ESS network
after the formal administrative structure of the nation that he imagined it
would network. “It is natural that the network should be supervised,” he
wrote, “by the Ministry of Communication [Svyaz’] in the Soviet Union,”
the ministry that managed many preexisting networks for information
exchange, including telegraph, telephone, phototelegraph, messages
(courier), and early digital technologies then available in small numbers.
Kharkevich breezily dismissed the distributed network model that Paul
Baran was developing at the time (although not by that name), observ-
ing that the structure of the network needs to be able to connect any two
nodes, and he writes, “in order to do this, it goes without saying, one does
not need to unite all nodes with separate lines.”^53 Instead, Kharkevich con-
sidered a hierarchically decentralized design, or pyramid structure, “the
rational structure of a network.”^54 Like transport roads, his network would
split out in a “radial system” in which a “given territorial group is united by
links to a communication node.” Just as every local, regional, and territorial
group would have its own common node, Kharkevich was quick to stress
the center that was implicit in this “radial” design. In 1962 in Moscow, a
“centralized automated management” design would have appeared reason-
able, if still monumental in aspiration, to him:

The brisk carrying out of these functions is possible only ... if the entire network will
work under centralized automated management. The governing center of ESS should
distribute information about the state of the network at every given moment....
The center should be capable of predicting such changes [in the network traffic]

Free download pdf